imgact_linux.c revision 119842
1/*-
2 * Copyright (c) 1994-1996 S�ren Schmidt
3 * All rights reserved.
4 *
5 * Based heavily on /sys/kern/imgact_aout.c which is:
6 * Copyright (c) 1993, David Greenman
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 *    notice, this list of conditions and the following disclaimer
13 *    in this position and unchanged.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 *    notice, this list of conditions and the following disclaimer in the
16 *    documentation and/or other materials provided with the distribution.
17 * 3. The name of the author may not be used to endorse or promote products
18 *    derived from this software without specific prior written permission
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
21 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
22 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
23 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
24 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
25 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
29 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30 */
31
32#include <sys/cdefs.h>
33__FBSDID("$FreeBSD: head/sys/i386/linux/imgact_linux.c 119842 2003-09-07 13:23:45Z bde $");
34
35#include <sys/param.h>
36#include <sys/systm.h>
37#include <sys/exec.h>
38#include <sys/imgact.h>
39#include <sys/imgact_aout.h>
40#include <sys/kernel.h>
41#include <sys/lock.h>
42#include <sys/mman.h>
43#include <sys/mutex.h>
44#include <sys/proc.h>
45#include <sys/resourcevar.h>
46#include <sys/vnode.h>
47
48#include <vm/vm.h>
49#include <vm/vm_kern.h>
50#include <vm/vm_param.h>
51#include <vm/pmap.h>
52#include <vm/vm_map.h>
53#include <vm/vm_extern.h>
54
55#include <i386/linux/linux.h>
56
57static int	exec_linux_imgact(struct image_params *iparams);
58
59static int
60exec_linux_imgact(struct image_params *imgp)
61{
62    const struct exec *a_out = (const struct exec *) imgp->image_header;
63    struct vmspace *vmspace;
64    vm_offset_t vmaddr;
65    unsigned long virtual_offset, file_offset;
66    vm_offset_t buffer;
67    unsigned long bss_size;
68    struct thread *td = curthread;
69    int error;
70
71    if (((a_out->a_magic >> 16) & 0xff) != 0x64)
72	return -1;
73
74    /*
75     * Set file/virtual offset based on a.out variant.
76     */
77    switch ((int)(a_out->a_magic & 0xffff)) {
78    case 0413:
79	virtual_offset = 0;
80	file_offset = 1024;
81	break;
82    case 0314:
83	virtual_offset = 4096;
84	file_offset = 0;
85	break;
86    default:
87	return (-1);
88    }
89    bss_size = round_page(a_out->a_bss);
90#ifdef DEBUG
91    printf("imgact: text: %08lx, data: %08lx, bss: %08lx\n",
92	(u_long)a_out->a_text, (u_long)a_out->a_data, bss_size);
93#endif
94
95    /*
96     * Check various fields in header for validity/bounds.
97     */
98    if (a_out->a_entry < virtual_offset ||
99	a_out->a_entry >= virtual_offset + a_out->a_text ||
100	a_out->a_text & PAGE_MASK || a_out->a_data & PAGE_MASK)
101	return (-1);
102
103    /* text + data can't exceed file size */
104    if (a_out->a_data + a_out->a_text > imgp->attr->va_size)
105	return (EFAULT);
106    /*
107     * text/data/bss must not exceed limits
108     */
109    mtx_assert(&Giant, MA_OWNED);
110    if (a_out->a_text > maxtsiz ||
111	a_out->a_data + bss_size > imgp->proc->p_rlimit[RLIMIT_DATA].rlim_cur)
112	return (ENOMEM);
113
114    VOP_UNLOCK(imgp->vp, 0, td);
115
116    /* copy in arguments and/or environment from old process */
117    error = exec_extract_strings(imgp);
118    if (error)
119	goto fail;
120
121    /*
122     * Destroy old process VM and create a new one (with a new stack)
123     */
124    exec_new_vmspace(imgp, &linux_sysvec);
125    vmspace = imgp->proc->p_vmspace;
126
127    /*
128     * Check if file_offset page aligned,.
129     * Currently we cannot handle misalinged file offsets,
130     * and so we read in the entire image (what a waste).
131     */
132    if (file_offset & PAGE_MASK) {
133#ifdef DEBUG
134	printf("imgact: Non page aligned binary %lu\n", file_offset);
135#endif
136	/*
137	 * Map text+data+bss read/write/execute
138	 */
139	vmaddr = virtual_offset;
140	error = vm_map_find(&vmspace->vm_map, NULL, 0, &vmaddr,
141			    a_out->a_text + a_out->a_data + bss_size, FALSE,
142			    VM_PROT_ALL, VM_PROT_ALL, 0);
143	if (error)
144	    goto fail;
145
146	error = vm_mmap(kernel_map, &buffer,
147			round_page(a_out->a_text + a_out->a_data + file_offset),
148			VM_PROT_READ, VM_PROT_READ, 0,
149			(caddr_t) imgp->vp, trunc_page(file_offset));
150	if (error)
151	    goto fail;
152
153	error = copyout((void *)(uintptr_t)(buffer + file_offset),
154			(void *)vmaddr, a_out->a_text + a_out->a_data);
155
156	vm_map_remove(kernel_map, buffer,
157		      buffer + round_page(a_out->a_text + a_out->a_data + file_offset));
158
159	if (error)
160	    goto fail;
161
162	/*
163	 * remove write enable on the 'text' part
164	 */
165	error = vm_map_protect(&vmspace->vm_map,
166			       vmaddr,
167			       vmaddr + a_out->a_text,
168			       VM_PROT_EXECUTE|VM_PROT_READ,
169			       TRUE);
170	if (error)
171	    goto fail;
172    }
173    else {
174#ifdef DEBUG
175	printf("imgact: Page aligned binary %lu\n", file_offset);
176#endif
177	/*
178	 * Map text+data read/execute
179	 */
180	vmaddr = virtual_offset;
181	error = vm_mmap(&vmspace->vm_map, &vmaddr,
182			a_out->a_text + a_out->a_data,
183			VM_PROT_READ | VM_PROT_EXECUTE,
184			VM_PROT_ALL,
185			MAP_PRIVATE | MAP_FIXED,
186			(caddr_t)imgp->vp, file_offset);
187	if (error)
188	    goto fail;
189
190#ifdef DEBUG
191	printf("imgact: startaddr=%08lx, length=%08lx\n",
192	    (u_long)vmaddr, a_out->a_text + a_out->a_data);
193#endif
194	/*
195	 * allow read/write of data
196	 */
197	error = vm_map_protect(&vmspace->vm_map,
198			       vmaddr + a_out->a_text,
199			       vmaddr + a_out->a_text + a_out->a_data,
200			       VM_PROT_ALL,
201			       FALSE);
202	if (error)
203	    goto fail;
204
205	/*
206	 * Allocate anon demand-zeroed area for uninitialized data
207	 */
208	if (bss_size != 0) {
209	    vmaddr = virtual_offset + a_out->a_text + a_out->a_data;
210	    error = vm_map_find(&vmspace->vm_map, NULL, 0, &vmaddr,
211				bss_size, FALSE, VM_PROT_ALL, VM_PROT_ALL, 0);
212	    if (error)
213		goto fail;
214#ifdef DEBUG
215	    printf("imgact: bssaddr=%08lx, length=%08lx\n",
216		(u_long)vmaddr, bss_size);
217#endif
218
219	}
220	/* Indicate that this file should not be modified */
221	mp_fixme("Unlocked v_flag access");
222	imgp->vp->v_vflag |= VV_TEXT;
223    }
224    /* Fill in process VM information */
225    vmspace->vm_tsize = round_page(a_out->a_text) >> PAGE_SHIFT;
226    vmspace->vm_dsize = round_page(a_out->a_data + bss_size) >> PAGE_SHIFT;
227    vmspace->vm_taddr = (caddr_t)(void *)(uintptr_t)virtual_offset;
228    vmspace->vm_daddr = (caddr_t)(void *)(uintptr_t)
229	(virtual_offset + a_out->a_text);
230
231    /* Fill in image_params */
232    imgp->interpreted = 0;
233    imgp->entry_addr = a_out->a_entry;
234
235    imgp->proc->p_sysent = &linux_sysvec;
236
237fail:
238    vn_lock(imgp->vp, LK_EXCLUSIVE | LK_RETRY, td);
239    return (error);
240}
241
242/*
243 * Tell kern_execve.c about it, with a little help from the linker.
244 */
245static struct execsw linux_execsw = { exec_linux_imgact, "linux a.out" };
246EXEC_SET(linuxaout, linux_execsw);
247